Our hypothesis was that MR microscopy might provide in vivo characterization of tissues in animals exposed to environmental toxins. We have developed techniques allowing us to follow serially the changes in T1 and T2 of diethylnitrosamine (DEN)-induced liver tumors in the same animal with resolution ~ 250 X that of earlier studies. This has permitted a quantitative measure in vivo of changes in tumor heterogeneity. The development of implanted rf coils has allowed us to extend microscopy to models of renal toxicity with pixels 2500 X smaller than early studies. Recently, we have demonstrated an improvement of > 50,000 X in resolution in in vitro samples by operating at ultrahigh (7.0 T) fields. We propose in this renewal to continue to these efforts by extending in vivo microscopy to pixels of 20 X 20 X 20 mum, i.e., 2 X 106 X greater than earlier studies. We will accomplish this through (a) operation at higher field (7 T), (b) use of implanted coils, and (c) further pulse sequence developments. Technical efforts will be directed at three specific models of liver toxicity; (a) CCl4-induced fibrosis, (b) bile-duct hyperplasia induced with alpha-naphthylisothiocyanate (ANIT), and (c) foci of cellular alteration produced by diethylnitrosamine (DEN). We will determine what elements in the MR microscopy images are relevant to the toxicologic pathologist. We will define the appropriate imaging techniques for in vivo MR microscopy in these models. With the improvements proposed here we will be able to follow the course of tissue changes in live animals at significantly higher resolution permitting for the first time in vivo histology.